X-MiND is an explainable framework for detecting mental-status shifts in timestamped user digital traces, such as social media posts, chats, and online interactions.
The name X-MiND stands for eXplainable Mental-status shift Detection.
The repository accompanies the paper:
Explainable Detection of Mental Status Shifts in User Digital Traces
Under revision at Social Network Analysis and Mining
Users generate digital traces that may reflect aspects of their mental state over time. X-MiND organizes these traces into temporal trajectories and analyzes them to identify phases of improvement, deterioration, or stability.
The framework combines:
- BERT-based models for extracting signals such as sentiment, emotion, and depression severity;
- temporal aggregation to build user-level trajectories;
- change-point detection to identify meaningful mental-status shifts;
- large language models to generate concise and human-readable reports.
X-MiND is intended for research purposes and does not aim to provide clinical diagnosis.
Digital traces
↓
BERT-based signal extraction
↓
Temporal trajectory construction
↓
Change-point detection
↓
LLM-based explainable report
git clone https://github.com/<username>/X-MiND.git
cd X-MiND
pip install -r requirements.txtpython src/main.py --input data/sample_data.csv --output results/Please adapt the command according to the final repository structure.
A minimal input file should contain timestamped user texts:
user_id,timestamp,text
If you use this repository, please cite:
@article{marozzo2026xmind,
title = {Explainable Detection of Mental Status Shifts in User Digital Traces},
author = {Marozzo, Fabrizio and others},
journal = {Social Network Analysis and Mining},
year = {2026},
note = {Under revision}
}Please update the citation with the final bibliographic details once the paper is accepted.
Fabrizio Marozzo
University of Calabria
Email: fabrizio.marozzo@unical.it